
基本信息:
- 专利标题: 一种用于链路预测的深度学习降维方法和装置
- 专利标题(英):Deep learning dimension-reduction method and device for link prediction
- 申请号:CN201710195501.3 申请日:2017-03-29
- 公开(公告)号:CN106972967A 公开(公告)日:2017-07-21
- 发明人: 高昕 , 张艳 , 李太松 , 邹潇湘 , 舒敏 , 云晓春 , 颜永红 , 张震 , 计哲 , 王锟 , 侯美佳 , 彭义刚 , 金暐 , 董琳
- 申请人: 国家计算机网络与信息安全管理中心 , 中国科学院声学研究所
- 申请人地址: 北京市朝阳区裕民路甲3号;
- 专利权人: 国家计算机网络与信息安全管理中心,中国科学院声学研究所
- 当前专利权人: 国家计算机网络与信息安全管理中心,中国科学院声学研究所
- 当前专利权人地址: 北京市朝阳区裕民路甲3号;
- 代理机构: 工业和信息化部电子专利中心
- 代理人: 于金平
- 主分类号: H04L12/24
- IPC分类号: H04L12/24 ; G06N99/00 ; G06Q10/04
The invention provides a deep learning dimension-reduction method and device of link prediction. The method comprises the following steps: determining a primary connecting network node and a secondary connecting network node of each network node according to a connecting relation of each network node in a set time slot; dividing the set time slot into multiple time slices according to the set time length, and determining a connecting relation of each network node and the corresponding primary connecting network node and the secondary connecting network node in each time slice according to the connecting relation of each network node in each time slice; performing link prediction on each network node through a deep learning algorithm model according to the connecting relation of each network node and the corresponding primary connecting network node and the secondary connecting network node in each time slice. By use of the method provided by the invention, the data volume input to the deep learning algorithm model is reduced, the learning training time is reduced, and the accuracy of the link prediction is improved.
公开/授权文献:
- CN106972967B 一种用于链路预测的深度学习降维方法和装置 公开/授权日:2020-07-24